Engineering of amine dehydrogenase for asymmetric reductive amination of ketone by evolving Rhodococcus phenylalanine dehydrogenase

Article abstract of DOI:10.1021/cs501906r

Triple mutant K66Q/S149G/N262C (TM-pheDH) of Rhodococcus phenylalanine dehydrogenase (pheDH) was engineered by directed evolution as the first enzyme for the highly enantioselective reductive amination of phenylacetone 1 and 4-phenyl-2-butanone 3, giving (R)-amphetamine 2 and (R)-1-methyl-3-phenylpropylamine 4 in >98% ee, respectively. The new amine dehydrogenase TM-pheDH with special substrate specificity is a valuable addition to the amine dehydrogenase family with very limited number, for asymmetric reductive amination of ketone, an important reaction in sustainable pharmaceutical manufacturing. Molecular docking provided insight into the role of key mutations of pheDH, being useful for engineering new amine dehydrogenases with higher activity and unique substrate scope. (Chemical Equation Presented).

Full text of DOI:10.1021/cs501906r

Letter
pubs.acs.org/acscatalysis
Engineering of Amine Dehydrogenase for Asymmetric Reductive
Amination of Ketone by Evolving Rhodococcus Phenylalanine
Dehydrogenase
Li Juan Ye, Hui Hung Toh, Yi Yang, Joseph P. Adams, Radka Snajdrova, and Zhi Li*
†
†
†
‡
‡
,†
†Department of Chemical and Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117585,
Singapore
‡
Medicines Research Centre, GlaxoSmithKline R&D Ltd, Gunnels Wood Road, Stevenage, Hertfordshire, SG1 2NY U.K.
*
S Supporting Information
ABSTRACT: Triple mutant K66Q/S149G/N262C
(
(
TM_pheDH) of Rhodococcus phenylalanine dehydrogenase
pheDH) was engineered by directed evolution as the first
enzyme for the highly enantioselective reductive amination of
phenylacetone 1 and 4-phenyl-2-butanone 3, giving (R)-
amphetamine 2 and (R)-1-methyl-3-phenylpropylamine 4 in
>
98% ee, respectively. The new amine dehydrogenase
TM_pheDH with special substrate specificity is a valuable
addition to the amine dehydrogenase family with very limited
number, for asymmetric reductive amination of ketone, an
important reaction in sustainable pharmaceutical manufacturing.
Molecular docking provided insight into the role of key mutations of pheDH, being useful for engineering new amine
dehydrogenases with higher activity and unique substrate scope.
KEYWORDS: amine dehydrogenase, biocatalysis, biotransformation, chiral amine, directed evolution, reductive amination
nantiopure amines are important building blocks for
pharmaceutical manufacturing. Enzyme catalysis provides
Scheme 1. Asymmetric Amination of Phenylacetone 1 and 4-
Phenyl-2-butanone 3 To Produce (R)-Amphetamine 2 and
(R)-1-Methyl-3-phenylpropylamine 4, Respectively, with
New Amine Dehydrogenase TM_pheDH Evolved from
Rhodococcus Phenylalanine Dehydrogenase
1
E
green and selective tools for the enantioselective syntheses of
2
,3
this class of chiral molecules. Many enzymes such as lipase,
4
,5
6,7
8,9
amine oxidase, imine reductase, transaminase, and amine
1
0−13
dehydrogenase
are reported for chiral amine synthesis.
Among them, amine dehydrogenase is attractive, because it
catalyzes the asymmetric reductive amination of ketone, a
highly wanted reaction in green and sustainable pharmaceutical
1
4
manufacturing, and it uses cheap ammonia as the reagent and
generates only water as byproduct, with high atom efficiency.
However, only one wild-type amine dehydrogenase (from
15
Streptomyces Sp) was reported, with low enantioselectivity.
Recently, directed evolution has become a useful tool for
engineering enzymes with new substrate acceptance and
16−18
improved catalytic performance.
By using this method,
Bommarius et al. engineered two new amine dehydrogenases
from Bacillus stereothermophilus leucine dehydrogenase and
Bacillus badius phenylalanine dehydrogenase for the preparation
methyl-3-phenylpropylamine 4, respectively, are selected as the
target reactions. (R)-Amphetamine 2 is a useful intermediate
for the preparation of (R, R)-formoterol, a potent bronchodi-
10−13
20
21
of several enantiopure amines.
To further expend the
lator, and tamsulosin, a prostate drug. (R)-1-Methyl-3-
phenylpropylamine 4 is a precursor of the antihypertensive
synthetic scope of amine dehydrogenase, we explored the
evolution of phenylalanine dehydrogenase (pheDH) from
22
dilevalol. The asymmetric synthesis of (R)-2 and (R)-4 via
amination of ketone 1 and 3 with ammonia is highly desirable,
19
Rhodococcus sp. M4 to create new amine dehydrogenases
with different substrate specificity for the asymmetric reductive
amination of ketones (Scheme 1).
Received: November 30, 2014
Revised: January 15, 2015
The asymmetric reduction of phenylacetone 1 and 4-phenyl-
2-butanone 3 to produce (R)-amphetamine 2 and (R)-1-
©
XXXX American Chemical Society
1119
DOI: 10.1021/cs501906r
ACS Catal. 2015, 5, 1119−1122

ACS Catalysis
Letter
but no amine dehydrogenases have been reported for these
reactions. Rhodococcus pheDH was selected as starting template
for the evolution, because the enzyme structure (PDB: 1C1D)
respectively. These activities are higher than those obtained
with other double mutants. In comparison with the wild-type
pheDH, all positive mutants contain relatively more hydro-
phobic amino acid residues at position 66 and 262 (S, M, Q
more hydrophobic than K, and C, F, I, L more hydrophobic
than N). Interestingly, the best double mutants K66Q/N262C
and K66Q/N262L are not among the reported 21 double
23,24
and catalytic mechanism
were reported, which provides a
solid basis for identifying a very limited number of key amino
acid residues for saturation mutagenesis; in addition, this
25
enzyme shares only 32% identity with Bacillus badius pheDH,
11
thus providing the potential to generate new amine
dehydrogenase with different substrate specificity than the
one engineered from Bacillus badius pheDH. In comparison
with Bacillus stereothermophilus leucine dehydrogenase, Rhodo-
coccus pheDH accepts a nature substrate which is structurally
more similar to the target ketone substrate 1 and 3.
As the targeted ketone substrates 1 and 3 do not contain a
carboxyl group, the amino acid residues Lys 66 and Asn262 of
pheDH, known to interact with the carboxyl group of the
natural substrate, were selected for simultaneous double-site
saturation mutagenesis. The saturation mutagenesis was
performed by PCR using NNK degenerate codons, the mutated
genes were transformed in E. coli, and the cells were grown on
agar plate. As the theoretical number of enzyme mutants is 400,
mutants at similar positions of pheDH from Bacillus badius.
To investigate the importance of single mutation at each of
the two selected positions, three single mutants K66Q, N262L,
and N262C were prepared. Mutant K66Q showed amination
activity toward both ketones 1 and 3, but the activity is
significantly lower than those with the two best double mutants
(Table 1). This suggests the importance of the K66Q mutation
27
and a synergetic effect of the double mutants. Two other
single mutants N262L and N262C did not show any activity for
the amination of 1 or 3. This further confirmed the mutation
K66Q as the key point of the success in changing substrate
acceptance from a keto acid to a ketone.
To further enhance the amination activity, we performed
another round of evolution with K66Q/N262C (DM_pheDH)
as starting enzyme. Twenty amino acid residues located within
4
230 clones were picked up for screening to ensure the
26
6
Å of phenylalanine bound in the catalytic center of the
coverage of possible mutants of >95%. A formazan-based
1
0
pheDH (Figure S2), except K78 and D118 which are essential
colorimetric assay was used to screen enzyme activity for the
deamination of racemic amine 2 and 4, respectively, with the
cell-free extract of E. coli expressing pheDH mutant. Forty-one
positive clones were identified, and six double mutants of
pheDH were confirmed with the deamination activity toward
amines 2 and 4, respectively. The reversed reactions,
asymmetric amination of ketone 1 and 3, were then
investigated with these double mutants, and the amine products
were analyzed by HPLC. As shown in Table 1, four double
mutants showed the amination activity for ketone 1 and four
double mutants for ketone 3. Among them, two double
mutants K66Q/N262L and K66Q/N262C accepted both
ketones 1 and 3 as substrates, giving amine 2 and 4 with a
specific activity of 3.9−3.6 and 5.2−6.2 U/g protein,
24
for the catalysis, were selected for single-site saturation
2
8,29
mutation.
The mutant library was built by using primers
containing NNK codons (Table S3). After transformation of
the plasmid in E. coli and cell growth on an agar plate, 3760
clones were picked up for screening with 95% coverage of the
possible mutants. A screening assay based on UV detection of
NADH formation in the deamination of racemic 2 and 4,
respectively, was used, and only one triple mutant K66Q/
S149G/N262C (TM_pheDH) was identified with higher
amination activity of 1 and 3 than the double mutant (Table
1
; 5.0 and 8.8 versus 3.6 and 6.2 U/g protein). The mutation
S149G replaced a hydrophilic amino acid by a hydrophobic
one.
The product ee of the asymmetric amination of ketone 1 and
3
, respectively, with the more active mutants K66Q/N262L,
K66Q/N262C (DM_pheDH) and K66Q/S149G/N262C
TM_pheDH), respectively, was checked by HPLC with a
Table 1. Directed Evolution of PheDH for Asymmetric
Amination of Phenylacetone 1 and 4-Phenyl-2-butanone 3,
respectively
(
chiral column. No peak of (S)-enantiomer existed in the chiral
HPLC chromatograms suggested >98% ee for (R)-2 and (R)-4,
respectively, for each of these three mutants. The product ee
values of other less active mutants were not determined.
To obtain insight into the role of the mutations on accepting
ketone 1 and 3 as the substrate and influencing the
enantioselectivity of the reductive aminations, the reported in
silico modeling and substrate docking method
establish the structure models of the mutants based on the
structure of pheDH and generate the active binding pose of
the substrate in the structure model of the mutants or structure
of pheDH. In the obtained active binding pose of phenyl
pyruvate in pheDH (Figure S4), the orientation of the substrate
resembles that of L-phenylalanine in the enzyme crystal
specific activi-
a
ty
U (g protein)−1
round no. of clones screened
positive mutants
1 to 2 3 to 4
b
1
4230
K66M/N262I
K66S/N262I
K66Q/N262I
K66Q/N262F
K66Q/N262L
K66Q/N262C
K66Q
0
0.4
0
1.7
0
3
0,31
2.8
0
were used to
3.2
3.9
3.6
1.1
0
24
5.2
6.2
0.9
0
c
2
N262L
N262C
0
0
d
3
3760
K66Q/S149G/N262C
5.0
8.8
24
structure, thus confirming the reliability of the modeling
method. In this pose, the carbonyl-O atom of phenyl pyruvate
formed hydrogen bond with Lys78. In comparison, the binding
poses of substrates 1 and 3 in pheDH (Figure 1a,b) were totally
different from the active conformation of natural substrate in
pheDH, and the carbonyl group of the substrate is far away
from Lys78 and Asp118 that are responsible for the formation
of imine intermediate based on the reported mechanism of
a
Reactions were performed in NH OH/NH Cl buffer (0.5 M; pH
.6) containing 5 mM substrate 1 or 3, 2 mM NADH, and cell-free
4
4
9
extract (1 g protein/L) of E. coli (pheDH mutant) at 30 °C and 250
rpm. Specific activity was determined for the first 30 min.
b
Simultaneous double-site saturation mutagenesis of K66 and N262.
c
d
Single mutation. Single-site saturation mutagenesis at 20 selected
amino acid residues.
1
120
DOI: 10.1021/cs501906r
ACS Catal. 2015, 5, 1119−1122

ACS Catalysis
Letter
−
4.3 kcal/mol for substrate 1 and −5.0 kcal/mol for substrate 3
in TM_pheDH and binding energy of −3.1 kcal/mol for both
substrate 3 and 4 in DM_pheDH. These resuls suggested lower
Km values of TM_pheDH and thus also higher catalytic
efficiency and specific activity for TM_pheDH.
The growth curve of E. coli (TM_pheDH) in TB medium
with the induction by IPTG is shown in Figure 2a. A cell
Figure 2. (a) Curves of cell growth of E. coli (TM_pheDH) in TB
medium at 22 °C and 250 rpm and specific enzyme activity for the
amination of 3 to (R)-4. Arrow indicates the addition of IPTG (0.5
mM) for the induction of TM_pheDH. (b) Time course of the
amination of 3 (15 mM) to (R)-4 with TM_pheDH (4 mg protein/
+
mL), GDH (40 U/mL), NAD (0.005 mM), and glucose (100 mM)
in NH OH/NH Cl buffer (0.5 M; pH 9.6) at 30 °C. (c) Scheme of
4
4
asymmetric amination of 4-phenyl-2-butanone 3 to (R)-4 with
cofactor recycling by using His-tagged TM_pheDH and GDH.
Figure 1. Enzyme−substrate binding pose for (a) wild type enzyme
with substrate 1, (b) wild type enzyme with substrate 3, (c) mutant
K66Q/N262C with substrate 1, (d) mutant K66Q/N262C with
substrate 3. (e) TM_pheDH with substrate 1, (f) TM_pheDH with
substrate 3. Mutated residues are shown in yellow. Distances (in
angstrom) are denoted by dashed lines.
density of around 4.5 g cdw/L was easily achieved at 9−11 h.
The amination activity of 3 with the cell-free extract of the cells
taken at different time points during growth was also shown in
Figure 2a. The highest specific activity (8.8 U/g protein) was
observed at 9 h in the late exponential growth phase.
24
pheDH for natural substrate. Such catalytically unfavorable
poses were caused by the formation of hydrogen bond between
the carbonyl-O atom of the substrates and Lys 66 (with a
distance of 3.2−3.3 Å). To make active enzyme for the desired
reductive amination of ketone, Lys 66 has to be mutated.
As shown in the poses of substrate 1 and 3 in the catalytic
pocket of DM_pheDH and TM_pheD (Figure 1c−f), no
binding between the carbonyl-O atom of the substrates and the
amino acid at 66 due to the K66Q mutation. The induction of
Asn262Cys mutation made the orientation of the terminal
methyl group of the substrate close to Cys262. The substrates
in the poses of Figure 1c−f showed similar position to that of
phenyl pyruvate in pheDH, with a distance of 2.9−3.1 Å
between substrate carbonyl-O atom and Lys78 due to hydrogen
bond. Based on the reported catalytic mechanism from the
natural substrate, Lys78 and the nearby Asp118 catalyze the
addition of ammonia to the carbonyl group of the substrate to
form the imine intermediate. A proton from NADH was then
added to the carbon atom of the imine group to give (R)-
configuration (Figure 1c−f), and further addition of a proton to
N atom gave amine (R)-2 or (R)-4.
His-tagged TM_pheDH was then produced by growing E.
coli cells expressing his-TM_PheDH and purified using a Ni-
NTA column. The kinetic data of TM_pheDH were obtained
for the asymmetric amination of ketone 1 and 3, respectively
(
Table 2). The enzyme has 3-fold higher catalytic efficiency
with substrate 3 than that with substrate 1. The k values are
0
cat
−
1
.70−0.72 s .
Asymmetric amination of 4-phenyl-2-butanone 3 to produce
(
R)-1-methyl-3-phenylpropylamine 4 with the recycling of
NADH was performed with TM_pheDH and glucose
32,33
dehydrogenase (GDH) (Figure 2b,c).
Reaction of 15
mM 3 gave 14.3 mM (R)- 4 at 60 h, with 95.2% conversion and
+
total turnover number of recycling NAD of 2800 (Figure 2b).
2
4
Table 2. Kinetic Data for the Asymmetric Amination of
Phenylacetone 1 and 4-Phenyl-2-butanone 3 with Purified
His-Tagged TM_pheDH, Respectively
a
Vmax mM min−1
kcat s−1
k /K s−1 mM−1
substrate
Km mM
cat
m
1
3
4.0
1.4
0.042
0.043
0.70
0.72
0.18
0.50
In comparison with DM_pheDH, TM_pheDH has the
additional S149G mutation which decreased side chain length
of the amino acide at 149 by ∼2 Å to enlarge the binding
pocket entrance, allowing for easier substrate access into the
binding pocket. Docking simulations gave binding energy of
a
Reaction was performed in NH OH/NH Cl buffer (0.5 M; pH 9.6)
4
4
containing 0.25−10 mM substrate 1 or 3, 2 mM NADH, and 1 μM
TM_pheDH at 30 °C for 10 min. Product concentration was
determined by HPLC analysis.
1
121
DOI: 10.1021/cs501906r
ACS Catal. 2015, 5, 1119−1122

ACS Catalysis
Letter
In summary, a triple mutant K66Q/S149G/N262C
TM_pheDH) was successfully engineered by directed
(8) Sehl, T.; Hailes, H. C.; Ward, J. M.; Wardenga, R.; von Lieres, E.;
Offermann, H.; Westphal, R.; Pohl, M.; Rother, D. Angew. Chem., Int.
Ed. 2013, 52, 6772−6775.
(
evolution of Rhodococcus pheDH via simultaneous random-
ization of two amino acid residues K66 and N262, followed by
single site saturation mutagenesis at other 20 selected residues,
as a new amine dehydrogenase for the highly enantioselective
reductive amination of phenylacetone 1 and 4-phenyl-2-
butanone 3 to give (R)-amphetamine 2 and (R)-1-methyl-3-
phenylpropylamine 4 in >98% ee, respectively. TM_pheDH
(
9) Kohls, H.; Steffen- Munsberg, F.; Hoehne, M. Curr. Opin Chem.
Biol. 2014, 19, 180−192.
10) Abrahamson, M. J.; Vazquez-Figueroa, E.; Woodall, N. B.;
Moore, J. C.; Bommarius, A. S. Angew. Chem., Int. Ed. 2012, 51, 3969−
972.
11) Abrahamson, M. J.; Wong, J. W.; Bommarius, A. S. Adv. Synth.
(
3
(
Catal. 2013, 355, 1780−1786.
−
1
showed a k of 0.70 and 0.72 s and k /K of 0.18 and 0.50
(12) Au, S. K.; Bommarius, B. R.; Bommarius, A. S. ACS catal. 2014,
4, 4021−4026.
cat
cat
m
−
1
−1
s
mM for the conversion of 1 to (R)-2 and of 3 to (R)-4,
(13) Bommarius, B. R.; Schurmann, M.; Bommarius, A. S. Chem.
respectively. Coupling of TM_pheDH and glucose dehydro-
Commun. 2014, 50, 14953−14955.
genase allowed 95% conversion of 15 mM 3 to (R)-4 with
+
(14) Constable, D. J. C.; Dunn, P. J.; Hayler, J. D.; Humphrey, G. R.;
Leazer, J. L.; Linderman, R. J.; Lorenz, K.; Manley, J.; Pearlman, B. A.;
Wells, A.; Zaks, A.; Zhang, T. Y. Green Chem. 2007, 9, 411−420.
NAD recycling. Molecular docking provided with some insight
into the role of key mutations, which could be very useful for
further engineering amine dehydrogenase with higher activity
and/or different substrate specificity. The engineered TM-
pheDH with different substrate specificity is an important
addition to the family of amine dehydrogenases with a very
limited number thus far, contributing to the expending of
synthetic scope of amine dehydrogenases-catalyzed asymmetric
reductive amination of ketone in green and sustainable
pharmaceutical manufacturing.
(15) Itoh, N.; Yachi, C.; Kukome, T. J. Mol. Catal. B: Enzym. 2000,
1
(
0, 281−290.
16) Turner, N. J. Nat. Chem. Biol. 2009, 5, 567−573.
(17) Auchli, R.; Rabe, K. S.; Kalbarczyk, K. Z.; Tata, A.; Heel, T.;
Kitto, R. Z.; Arnold, F. H. Angew. Chem., Int. Ed. 2013, 52, 5571−5574.
(18) Reetz, M. T. J. Am. Chem. Soc. 2013, 135, 12480−12496.
(19) Hummel, W.; Schutte, H.; Schmidt, E.; Wandrey, C.; Kula, M.
R. Appl. Microbiol. Biotechnol. 1987, 26, 409−416.
(20) Murase, K.; Mase, T.; Ida, H.; Takahashi, K.; Murakami, M.
Chem. Pharm. Bull. 1978, 26, 1123−1129.
(21) Abrams, P.; Speakman, M.; Stott, M.; Arkell, D.; Pocock, R. Br.
J. Urol. 1997, 80, 587−596.
ASSOCIATED CONTENT
■
*
S
Supporting Information
The following file is available free of charge on the ACS
Publications website at DOI: 10.1021/cs501906r.
(
22) Clifton, J. E.; Collins, I.; Hallett, P.; Hartley, D.; Lunts, L. H. C.;
Wicks, P. D. J. Med. Chem. 1982, 25, 670−679.
23) Vanhooke, J. L.; Thoden, J. B.; Brunhuber, N. M. W.; Blanchard,
(
Chemicals, strains and biochemicals, analytic method;
procedures for directed evolution, cell growth, enzyme
purification, biotransformation, enzyme simulation and
modeling, kinetics; HPLC chromatograms (PDF)
J. S.; Holden, H. M. Biochemistry 1999, 38, 2326−2339.
(24) Brunhuber, N. M. W.; Thoden, J. B.; Blanchard, J. S.; Vanhooke,
J. L. Biochemistry 2000, 39, 9174−9187.
(25) Yamada, A.; Dairi, T.; Ohno, Y.; Huang, X. L.; Asano, Y. Biosci.
Biotechnol. Biochem. 1995, 59, 1994−1995.
(26) Reetz, M. T.; Kahakeaw, D.; Lohmer, R. ChemBioChem. 2008, 9,
AUTHOR INFORMATION
■
*
6
1797−1804.
Corresponding Author
E-mail: chelz@nus.edu.sg. Phone: +65-6516 8416. Fax: +65-
779 1936.
(27) Reetz, M. T. Angew. Chem., Int. Ed. 2013, 52, 2658−2666.
(
28) Yang, Y.; Liu, J.; Li, Z. Angew. Chem., Int. Ed. 2014, 53, 3120−
3
124.
(29) Pham, S. Q.; Pompidor, G.; Liu, J.; Li, X. D.; Li, Z. Chem.
Notes
Commun. 2012, 48, 4618−4620.
(30) Kille, S.; Zilly, F. E.; Acevedo, J. P.; Reetz, M. Nat. Chem. 2011,
The authors declare no competing financial interest.
3
, 738−743.
31) Zhang, Z. G.; Lonsdale, R.; Sanchis, J.; Reetz, M. T. J. Am. Chem.
Soc. 2014, 136, 17262−17272.
32) Zhang, W.; O’Connor, K.; Wang, D. I. C.; Li, Z. Appl. Environ.
(
ACKNOWLEDGMENTS
■
(
This work was financially supported by GlaxoSmithKline
GSK) and Singapore Economic Development Board (EDB)
through a Green and Sustainable Manufacturing Grant (Project
No. 279-000-348-592).
(
Microbiol. 2009, 75, 687−694.
(33) Pham, S. Q.; Gao, P.; Li, Z. Biotechnol. Bioeng. 2013, 110, 363−
373.
REFERENCES
■
(
(
1) Ho
̈
hne, M.; Bornscheuer, U. T. ChemCatChem 2009, 1, 42−51.
2) Messina, F.; Botta, M.; Corelli, F.; Schneider, M. P.; Fazio, F. J.
Org. Chem. 1999, 64, 3767−3769.
3) Paetzold, J.; Backvall, J. E. J. Am. Chem. Soc. 2005, 127, 17620−
7621.
4) Li, T.; Liang, J.; Ambrogelly, A.; Brennan, T.; Gloor, G.;
(
1
(
Huisman, G.; Lalonde, J.; Lekhal, A.; Mijts, B.; Muley, S.; Newman, L.;
Tobin, M.; Wong, G.; Zaks, A.; Zhang, X. J. Am. Chem. Soc. 2012, 134,
6
(
(
467−6472.
5) Ghislieri, D.; Turner, N. J. Top Catal. 2013, 57, 284−300.
6) Leipold, F.; Hussain, S.; Ghislieri, D.; Turner, N. J.
ChemCatChem 2013, 5, 3505−3508.
7) Rodríguez-Mata, M.; Frank, A.; Wells, E.; Leipold, F.; Turner, N.
J.; Hart, S.; Turkenburg, J. P.; Grogan, G. ChemBioChem 2013, 14,
372−1379.
(
1
1
122
DOI: 10.1021/cs501906r
ACS Catal. 2015, 5, 1119−1122